Pith. sign in

REVIEW 3 major objections 4 minor 35 references

Multimodal Prescriptive Deep Learning

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper proposes Prescriptive Neural Networks (PNNs), a feedforward neural network that learns an outcome-optimizing treatment policy from multimodal data—tabular features plus text embeddings—and reports estimated complication…

desk verdict Useful but self-referential: headline gains are measured on the same estimated reward matrix used to train the policies, so the 32% and 40% numbers are not real-world evidence until validated. read the letter →

arxiv 2501.14152 v1 pith:6AOLYR4L submitted 2025-01-24 cs.LG stat.ML

Dimitris Bertsimas , Lisa Everest , Vasiliki Stoumpou This is my paper · ORCID
classification cs.LGstat.ML
keywords prescriptiveanalyticsmultimodaldeeplearningcounterfactualestimationdoublyrobustestimatorneuralnetworkpolicyknowledgedistillationoptimalclassificationtreeshealthcaredecision-making
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper's central claim is that a feedforward neural network can be trained to prescribe treatments by minimizing an estimated counterfactual reward, and that this framework is the first prescriptive method able to consume multimodal data. PNNs take tabular features and embeddings from pretrained clinical language models, estimate a doubly robust reward matrix over all treatments, and then train a softmax output layer to pick, for each patient, the treatment with the best estimated outcome. On two real-world medical datasets, the authors report that multimodal PNNs reduce estimated postoperative complication rates by 32% in TAVR and estimated mortality by over 40% in liver trauma. On four tabular datasets spanning binary, discrete, single-continuous, and multiple-continuous treatments, PNNs outperform or match established prescriptive baselines. The paper also shows that these neural prescriptions can be distilled into interpretable Optimal Classification Trees with little loss, which matters for deployment where decisions must be explained.

What carries the argument

The load-bearing object is the Prescriptive Neural Network (PNN), a feedforward network whose output layer has one neuron per treatment and uses a softmax to turn output logits into prescription probabilities. During training it minimizes the relaxed objective $\frac{1}{n}\sum_{i=1}^n \sum_{t\in T} \sigma_t(z(x_i))\Gamma_{i,t}$, where $\Gamma_{i,t}$ is the doubly robust estimated outcome of assigning treatment $t$ to observation $i$. This objective is the soft relaxation of the policy objective proposed by Amram et al. (2022), and it is what makes the prescription task differentiable and trainable by backpropagation. Multimodality enters through the input representation: embeddings extracted from Clinical Longformer text models are reduced by PCA and concatenated with normalized tabular features. Interpretability is then recovered by fitting an Optimal Classification Tree to the PNN's final prescriptions, yielding a Mirrored OCT.

What would settle it

Give the same PNN pipeline a synthetic dataset whose true outcome functions are known, and compare the expected true outcome of its prescriptions with the improvement estimated from the reward matrix; alternatively, run a prospective study in which TAVR and liver trauma patients are treated according to PNN prescriptions and compare actual complication and mortality rates with historical practice. If the synthetic or prospective true outcomes show no improvement, the central claim fails.

Watch

Extended reading notes

Core claim

The discovery claimed is that the non-differentiable prescriptive objective $\min_\tau \sum_i \mathbf{1}\{\tau(x_i)=t\}\Gamma_{i,t}$ can be replaced by the softmax-relaxed objective $\min_\tau \frac{1}{n}\sum_i \sum_t \sigma_t(z(x_i))\Gamma_{i,t}$, making the prescription problem trainable end-to-end with backpropagation. Here $\Gamma_{i,t}$ is the doubly robust estimate of the outcome under treatment $t$ for observation $i$, and $\sigma_t(z)$ is the softmax probability the network assigns to treatment $t$. With embeddings from Clinical Longformer concatenated to tabular features, the same network handles multimodal inputs. The authors claim this is the first prescriptive method for multimodal data, and that the resulting policies improve estimated outcomes by 32% and over 40% in the TAVR and liver trauma datasets while matching or beating four established baselines on tabular data.

Load-bearing premise

The load-bearing premise is that the estimated counterfactual reward matrix $\Gamma$, built from historical observational data with doubly robust estimation, accurately represents what each patient's outcome would actually have been under every treatment; if that matrix is biased, the reported 32% and 40% improvements and the favorable comparisons against baselines may not reflect real-world gains.

Editorial extensions

If this is right

  • Incorporating clinical notes alongside tabular features improved estimated outcome gains by more than 10 percentage points on average in the TAVR and liver trauma datasets, suggesting multimodal information is what drives the largest gains.
  • Across diabetes, groceries, splenic injuries, and REBOA datasets, PNNs matched or beat Regress & Compare, Causal Forests, and Optimal Policy Trees on estimated outcome improvement.
  • Fitting Mirrored OCTs to PNN prescriptions preserves most of the estimated improvement, with an average loss of only 1.38% across tabular datasets, while returning an interpretable decision tree.
  • PNNs prescribe a comparable share of the available treatments and remain stable across randomized splits, with standard deviations similar to the other prescriptive methods.
  • Because the network's final layer outputs a probability for each treatment, users can apply thresholds or penalties to control how much the prescribed policy deviates from current practice.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the prescription problem is framed as softmax classification over treatments, the same architecture should transfer to other multimodal decision settings—dynamic pricing with text reviews, ad targeting with images or video, and personalized maintenance scheduling—without changing the model.
  • The use of frozen pretrained embeddings means the framework could act as a lightweight prescriptive head on top of any large language or vision model, pointing toward a general recipe for turning generative models into decision policies.
  • Distilling a black-box policy into an OCT also provides an audit mechanism: inspecting the tree's splits could reveal whether the learned policy depends on clinically or ethically questionable features, helping to certify the policy before deployment.
  • Because the paper's evaluation metric uses the same reward matrix that defines the training target, the most informative next experiment is external validation—synthetic ground-truth counterfactuals or a prospective comparison of PNN prescriptions against clinician decisions.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper introduces Prescriptive Neural Networks (PNNs), a feedforward neural network trained on multimodal (tabular and text) embeddings to output outcome-optimizing treatment prescriptions. The training objective (Eq. 4) minimizes a softmax-weighted sum of entries from a counterfactual reward matrix Γ estimated by doubly robust or direct methods (Section 2.3). The authors report that PNNs reduce estimated postoperative complication rates by 32% in TAVR and estimated mortality by over 40% in liver trauma, and that on four tabular datasets PNNs outperform or match Regress & Compare, Causal Forests, and Optimal Policy Trees. They also fit Mirrored Optimal Classification Trees on PNN prescriptions to recover interpretability, and report stability and realism metrics across 5 data splits and 5 model runs.

Significance. If the empirical claims were reliable, the paper would make a useful contribution: it is a flexible prescriptive framework that handles multimodal inputs, spans all four treatment scenarios (binary, multiple discrete, single continuous, multiple continuous), and offers an interpretability recovery route via knowledge distillation. The stability and treatment-realism analyses are thoughtful additions. However, the central empirical evidence rests entirely on estimated counterfactual rewards, with no synthetic ground-truth validation, no observed-outcome calibration, and no formal statement or test of the unconfoundedness and positivity assumptions required by the doubly robust estimator. The 32% and 40% headline gains are therefore not established as real-world effects. The paper is internally consistent in its modeling, but its external validity claims are substantially overreaching.

major comments (3)
  1. [Section 3.4, Eq. (5), and Section 2.4, Eq. (4)] The evaluation metric in Eq. (5) is computed from the same estimated reward matrix Γ that appears in the training objective Eq. (4). A policy minimizing Eq. (4) will preferentially choose treatments whose Γ entries are underestimated, and Eq. (5) will mechanically report those gains. The paper acknowledges in Section 3.4 that 'test-set reward matrix entries lack natural meaning' for doubly robust estimates, yet the abstract and Section 1.2 claim 32% and 40% improvements in estimated outcomes as the paper's headline results. This circularity means the absolute improvements in Table 1 are not credible evidence of real-world benefit. The authors should validate Γ against either synthetic ground truth with known counterfactuals or observed factual outcomes in an overlap region, and should state and test unconfoundedness and positivity assumptions for Eq. (1). Without such validation, the absolute improvement claims should be explicitly framed as 'improvements in estimated rewards under the chosen estimator,' and the abstract should be tempered accordingly.
  2. [Section 3.1 and Section 3.5, Table 1] The procedure for estimating the test-set reward matrix Γ is underspecified. It is not stated whether the test-set counterfactuals are estimated with models trained only on the 50% training split (out-of-fold) or on the full data including the test set. If the latter, the evaluation leaks information and the reported improvement numbers, including the multimodal-versus-tabular comparisons in Table 1, are optimistically biased. This concern affects all results, not only the absolute numbers, and should be clarified and corrected if needed.
  3. [Section 3.5, Table 1] The claim that multimodal PNNs significantly outperform tabular-only PNNs is based on reward matrices that, in the 'Tabular & Notes' rows, are themselves estimated using the multimodal data. Since the PNN is trained on the same embedding representations used by the reward estimator, the multimodal advantage may partly reflect shared representation bias rather than genuine prescriptive gains from additional modalities. The paper reports results under both tabular-only and multimodal reward estimators, which is helpful, but the central multimodal claim would be considerably strengthened by a synthetic experiment where the true counterfactual outcomes are known, or by a factual-outcome validation under an overlap assumption.
minor comments (4)
  1. [Throughout] The manuscript contains several typos and formatting issues, including 'replcement' in Section 3.5, 'V asiliki' in the author line, and 'T ranscatheter aortic valve replcement' in Section 3.5. A careful proofread is needed.
  2. [Section 3.2] The hyperparameter choices (number of layers, nodes, learning rate, weight decay, batch size) are described qualitatively but no actual values are reported for any dataset. A table listing the tuned hyperparameters per dataset would substantially improve reproducibility.
  3. [Table 4, footnote 1] The revenue improvement formula for groceries uses the notation '¯pr' twice for different quantities; this should be disambiguated (e.g., one is the mean revenue under prescribed treatments and the other is the mean revenue under actual treatments).
  4. [Appendix A.2] Several figures referenced in the appendix (e.g., Figures 3-12) appear in the manuscript as 'Force Appendix above figures' placeholder text rather than as embedded figures; readers are directed to an external Google Drive link. The manuscript should either include the figures or clearly state that they are available as supplementary material.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline gains (32%/40%) are computed from the same estimated reward matrix Γ used to train the PNN, making the evaluation self-referential.

  1. fitted input called prediction [Section 2.4 Eq (4); Section 3.4 Eq (5); Section 3.5 Table 1]
    "The tractable objective for our PNN models is therefore: min τ(.) 1/n ∑_i ∑_t P[τ(x_i)=t]· Γ_i,t. (4) ... The average relative outcome improvement is then computed as: ¯I = ∑_i |Γ_i,ˆt_i − Γ_i,t_i| / ∑_i Γ_i,t_i. (5)"

    The PNN is trained by minimizing the Γ-weighted prescription objective (Eq 4). The headline improvement (Eq 5) measures the difference between the estimated outcomes Γ_i,ˆt_i of the PNN's prescriptions and the estimated outcomes Γ_i,t_i of the real-life treatments, using the same reward matrix Γ (or the same doubly robust estimator) that defined the training objective. Section 3.4 concedes that 'test-set reward matrix entries lack natural meaning' and therefore 'hypothetical outcomes for both treatments ... are drawn from the reward matrix and then compared.' Thus the reported 32%/40% reductions are reductions in the very estimated quantity the network was optimized against, not in observed or externally validated outcomes.

full rationale

The paper's central claims are the 32% and 40% reductions in estimated outcomes. The training objective (Eq 4) and the evaluation metric (Eq 5) both operate on Γ, the doubly robust reward matrix (Eq 1). The paper does not validate Γ against observed outcomes, synthetic ground truth, or prospective data, and does not state or test the unconfoundedness/positivity conditions required for Eq 1 to recover true counterfactuals. However, the rank comparisons against other prescriptive methods in Table 4 are less vulnerable because all methods are scored on the same Γ; those comparisons provide independent content. The vulnerable claims are the absolute improvements (32%/40%) and the multimodal advantage, which are self-referential: the network is trained to minimize Γ and evaluated on Γ. No other circular steps (e.g., load-bearing self-citations) were found; citations to Bertsimas & Dunn (OCT), Bertsimas & Kallus (R&C), etc. are standard tool citations and not load-bearing for the central claim. Score 6 reflects partial circularity: the headline predictions reduce to the same estimated input they were fitted on, but the train/test split and benchmark comparisons give some independent content.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The method relies on the estimated reward matrix as both training target and evaluation metric, plus several unreported tuning choices including embedding dimension, discretization grids, architecture, and estimator selection. The unconfoundedness assumption is not stated.

free parameters (6)
  • PCA embedding dimension = 32
    Clinical Longformer embeddings are reduced to 32 dimensions with no sensitivity analysis or justification for this choice (Section 2.2.2).
  • Continuous treatment discretization grid = not reported
    Continuous doses and prices are discretized into a finite subset of treatments; grid size and subset selection are not specified (Sections 2.3.2 and 3.6).
  • PNN hyperparameters = not reported
    Layers, nodes, batch size, learning rate, weight decay, and epochs are tuned per dataset on a validation set (Section 3.2), but final values are omitted.
  • Strawberry price grid = $2 to $5 in $0.50 increments
    The price grid is fixed for the groceries application with no justification for endpoints or step size (Section 3.6).
  • OCT maximum depth = 7
    Mirrored OCTs are displayed at maximum depth 7; the depth selection procedure is not described (Appendix A.2).
  • Counterfactual estimator choice per dataset = XGBoost or Random Forest, direct or doubly robust
    Estimator family and direct versus doubly robust mode vary by dataset (Table 3), with no comparison or sensitivity analysis.
assumptions (5)
  • domain assumption Unconfoundedness: treatment assignment is independent of potential outcomes given observed features.
    The doubly robust estimator corrects selection on observables but cannot handle unobserved confounding. This assumption is not stated or tested in Section 2.3.
  • domain assumption The doubly robust estimator yields approximately unbiased counterfactual outcome estimates Gamma.
    Both the training objective Eq (4) and the improvement metric Eq (5) treat Gamma as the true outcome. Section 4 states that estimated rewards are a good approximation without evidence.
  • ad hoc to paper Estimated reward improvement on a held-out split is a valid proxy for real clinical or business improvement.
    This is the central evaluation assumption. No synthetic ground-truth benchmark, prospective validation, or comparison to observed factual outcomes is provided.
  • domain assumption Clinical Longformer embeddings capture clinically relevant information.
    Embeddings are used as inputs; if they are uninformative, multimodal gains could reflect noise or overfitting (Section 2.2.2).
  • domain assumption The 50/50 split provides enough data for reliable test-set counterfactual estimation.
    Section 3.1 justifies the split by the need to estimate counterfactuals on the test set, but no diagnostic for estimator quality is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multimodal Prescriptive Deep Learning." pith.science (2026). https://pith.science/paper/6AOLYR4L

@misc{pith2026250114152,
  author       = {Pith},
  title        = {Pith review of: Multimodal Prescriptive Deep Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6AOLYR4L}},
  note         = {Machine review of arXiv:2501.14152}
}
read the original abstract

We introduce a multimodal deep learning framework, Prescriptive Neural Networks (PNNs), that combines ideas from optimization and machine learning, and is, to the best of our knowledge, the first prescriptive method to handle multimodal data. The PNN is a feedforward neural network trained on embeddings to output an outcome-optimizing prescription. In two real-world multimodal datasets, we demonstrate that PNNs prescribe treatments that are able to significantly improve estimated outcomes in transcatheter aortic valve replacement (TAVR) procedures by reducing estimated postoperative complication rates by 32% and in liver trauma injuries by reducing estimated mortality rates by over 40%. In four real-world, unimodal tabular datasets, we demonstrate that PNNs outperform or perform comparably to other well-known, state-of-the-art prescriptive models; importantly, on tabular datasets, we also recover interpretability through knowledge distillation, fitting interpretable Optimal Classification Tree models onto the PNN prescriptions as classification targets, which is critical for many real-world applications. Finally, we demonstrate that our multimodal PNN models achieve stability across randomized data splits comparable to other prescriptive methods and produce realistic prescriptions across the different datasets.

Figures

Figures reproduced from arXiv: 2501.14152 by the authors.

Figure 1
Figure 1. Architecture of a feedforward neural network (Quiza & Davim (2011)). [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Example of REBOA Mirrored OCT. 3 Experiments with real-world datasets In this section, we apply PNNs on real-world datasets that are both multimodal and unimodal. We first review methodology for data splits, network architecture, and performance evaluation, which are relevant for all of our experiments. We then report results and relevant discussions for each of our two multimodal datasets and four unimodal datasets… view at source ↗
Figure 3
Figure 3. Mirrored OCT of maximum depth 7 for TAVR. [PITH_FULL_IMAGE:figures/full_fig_p022_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Mirrored OCT of maximum depth 7 for liver trauma. [PITH_FULL_IMAGE:figures/full_fig_p023_4.png]
Figure 5
Figure 5. Figure 5: Mirrored OCT of maximum depth 7 for diabetes management. [PITH_FULL_IMAGE:figures/full_fig_p024_5.png]
Figure 6
Figure 6. Figure 6: Optimal Policy Tree of maximum depth 7 for diabetes management. [PITH_FULL_IMAGE:figures/full_fig_p025_6.png]
Figure 7
Figure 7. Figure 7: Mirrored OCT of maximum depth 7 for groceries. [PITH_FULL_IMAGE:figures/full_fig_p026_7.png]
Figure 8
Figure 8. Figure 8: Optimal Policy Tree of maximum depth 7 for groceries. [PITH_FULL_IMAGE:figures/full_fig_p027_8.png]
Figure 9
Figure 9. Figure 9: Mirrored OCT of maximum depth 7 for spleen injury treatment. [PITH_FULL_IMAGE:figures/full_fig_p028_9.png]
Figure 10
Figure 10. Figure 10: Optimal Policy Tree of maximum depth 7 for spleen injury treatment. [PITH_FULL_IMAGE:figures/full_fig_p029_10.png]
Figure 11
Figure 11. Figure 11: Mirrored OCT of maximum depth 7 for REBOA in blunt trauma patients. [PITH_FULL_IMAGE:figures/full_fig_p030_11.png]
Figure 12
Figure 12. Figure 12: Optimal Policy Tree of maximum depth 7 for REBOA in blunt trauma patients. [PITH_FULL_IMAGE:figures/full_fig_p031_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 25 canonical work pages

  1. [2]

    Optimal policy trees

    Maxime Amram, Jack Dunn, and Ying Daisy Zhuo. Optimal policy trees. Machine Learning, 11: 0 2741--2768, 2022

    work page 2022

  2. [3]

    Recursive partitioning for heterogeneous causal effects

    Susan Athey and Guido Imbens. Recursive partitioning for heterogeneous causal effects. Proceedings of the National Academy of Sciences, 113 0 (27): 0 7353--7360, 2016. URL https://doi.org/10.1073/pnas.1510489113

    doi:10.1073/pnas.1510489113 2016

  3. [4]

    Mack, George Ruiz, Mark S

    Mohsen Bayati, Mark Braverman, Michael Gillam, Karen M. Mack, George Ruiz, Mark S. Smith, and Eric Horvitz. Data-driven decisions for reducing readmissions for heart failure: General methodology and case study. PLOS ONE, 9(10), 2014. URL https://doi.org/10.1371/journal.pone.0109264

    doi:10.1371/journal.pone.0109264 2014

  4. [5]

    Janos: an integrated predictive and prescriptive modeling framework

    David Bergman, Teng Huang, Philip Brooks, Andrea Lodi, and Arvind U Raghunathan. Janos: an integrated predictive and prescriptive modeling framework. INFORMS Journal on Computing, 34 0 (2): 0 807--816, 2022

    work page 2022

  5. [6]

    Optimal classification trees

    Dimitris Bertsimas and Jack Dunn. Optimal classification trees. Machine Learning, 106: 0 1039–1082, 2017

    work page 2017

  6. [7]

    Machine Learning under a Modern Optimization Lens

    Dimitris Bertsimas and Jack Dunn. Machine Learning under a Modern Optimization Lens. Dynamic Ideas, 2019

    work page 2019

  7. [8]

    From predictive to prescriptive analytics

    Dimitris Bertsimas and Nathan Kallus. From predictive to prescriptive analytics. Management Science, 66(3):1025-1044, 2020. URL https://pubsonline.informs.org/doi/10.1287/mnsc.2018.3253

    arXiv 2020

  8. [9]

    The power and limits of predictive approaches to observational data-driven optimization: The case of pricing

    Dimitris Bertsimas and Nathan Kallus. The power and limits of predictive approaches to observational data-driven optimization: The case of pricing. INFORMS Journal on Optimization, 5 0 (1), 2022. URL https://www.researchgate.net/publication/302568371_Pricing_from_Observational_Data

    work page arXiv 2022

Show all 35 references
  1. [10]

    Personalized diabetes management using electronic medical records

    Dimitris Bertsimas, Nathan Kallus, Alexander M Weinstein, and Daisy Ying Zhuo. Personalized diabetes management using electronic medical records. Diabetes care, 40(2):210–217, 2017 a . URL https://pubmed.ncbi.nlm.nih.gov/27920019/

    2017

  2. [11]

    From predictive methods to missing data imputation: An optimization approach

    Dimitris Bertsimas, Colin Pawlowski, and Ying Daisy Zhuo. From predictive methods to missing data imputation: An optimization approach. J. Mach. Learn. Res., 18: 0 196:1--196:39, 2017 b

    2017

  3. [12]

    Optimal prescriptive trees

    Dimitris Bertsimas, Jack Dunn, and Nishanth Mundru. Optimal prescriptive trees. INFORMS Journal on Optimization, 1(2):164–183, 2019. URL https://jack.dunn.nz/papers/OptimalPrescriptiveTrees.pdf

    2019

  4. [13]

    Model distillation for revenue optimization: Interpretable personalized pricing, 2021

    Max Biggs, Wei Sun, and Markus Ettl. Model distillation for revenue optimization: Interpretable personalized pricing, 2021

    2021

  5. [15]

    Classification and regression trees

    Leo Breiman, Jerome Friedman, Charles Stone, and Richard Olshen. Classification and regression trees. CRC press, 1984

    1984

  6. [16]

    Cleveland and Susan J

    William S. Cleveland and Susan J. Devlin. Locally weighted regression: An approach to regression analysis by local fitting. Journal of the American Statistical Association, 83(403):596-610, 1988. URL https://www.jstor.org/stable/2685209

    1988

  7. [17]

    The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and smale’s 18th problem

    Matthew J Colbrook, Vegard Antun, and Anders C Hansen. The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and smale’s 18th problem. Proceedings of the National Academy of Sciences, 119 0 (12): 0 e2107151119, 2022

    2022

  8. [18]

    Doubly robust policy evaluation and learning

    Miroslav Dudik, John Langford, and Lihong Li. Doubly robust policy evaluation and learning. arXiv preprint arXiv:11034601, 2011. URL https://arxiv.org/abs/1103.4601

    2011 arXiv

  9. [19]

    Causal modeling and inference for electricity markets

    Egil Ferkingstad, Anders Løland, and Mathilde Wilhelmsen. Causal modeling and inference for electricity markets. Energy Economics, 33(3):404-412, 2011. URL https://doi.org/10.1016/j.eneco.2010.10.006

    2011 doi

  10. [20]

    Accuracy of the aast organ injury scale for ct evaluation of traumatic liver and spleen injuries

    Homann Georg, Toschke Christina, Gassmann Peter, and Vieth Volker. Accuracy of the aast organ injury scale for ct evaluation of traumatic liver and spleen injuries. Chinese Journal of Traumatology, 17 0 (01): 0 25--30, 2014

    2014

  11. [21]

    Recursive partitioning for personalization using observational data

    Nathan Kallus. Recursive partitioning for personalization using observational data. Proceedings of Machine Learning Research, 70:1789-1798, 2017. URL https://proceedings.mlr.press/v70/kallus17a.html

    2017

  12. [22]

    Prosthesis tailoring for patients undergoing transcatheter aortic valve implantation

    Pier Pasquale Leone, Andrea Scotti, Edwin C Ho, Manaf Assafin, James Doolittle, Mei Chau, Leandro Slipczuk, Matthew Levitus, Damiano Regazzoli, Antonio Mangieri, et al. Prosthesis tailoring for patients undergoing transcatheter aortic valve implantation. Journal of Clinical Me...

    2023

  13. [23]

    Wehbe, Faraz S

    Yikuan Li, Ramsey M. Wehbe, Faraz S. Ahmad, Hanyin Wang, and Yuan Luo. Clinical-longformer and clinical-bigbird: Transformers for long clinical sequences, 2022. URL https://arxiv.org/abs/2201.11838

    2022 arXiv

  14. [24]

    The complete journey, 2020

    Steven Lugauer, Steven Buechler, Timothy Gilbride, and Carrie Heilman. The complete journey, 2020. Data retrieved from dunnhumby, http://www.dunnhumby.com/source-files/

    2020

  15. [25]

    Nienaber

    Andreas Mitsis, Xun Yuan, Christos Eftychiou, Panayiotis Avraamides, and Christoph A. Nienaber. Personalised treatment in aortic stenosis: A patient-tailored transcatheter aortic valve implantation approach. Journal of Cardiovascular Development and Disease, 9 0 (11), 2022. IS...

    2022 doi

  16. [26]

    Applications of 0-1 neural networks in prescription and prediction

    Vrishabh Patil, Kara Hoppe, and Yonatan Mintz. Applications of 0-1 neural networks in prescription and prediction. A key challenge in medical decision making is learning treatment policies for patients with limited observational data. This challenge is particularly evident in ...

    2024 arXiv

  17. [27]

    Shah, Trevor Hastie, and Robert Tibshirani

    Scott Powers, Junyang Qian, Kenneth Jung, Alejandro Schuler, Nigam H. Shah, Trevor Hastie, and Robert Tibshirani. Some methods for heterogeneous treatment effect estimation in high dimensions. Statistics in Medicine, 37 0 (11): 0 1767–1787, 2018. doi:10.1002/sim.7623

    2018 doi

  18. [28]

    Ramon Quiza and J. Davim. Computational Methods and Optimization, pp.\ 177--208. Springer-Verlag, 01 2011. ISBN 978-1-84996-449-4. doi:10.1007/978-1-84996-450-0

    2011 doi

  19. [29]

    Causal inference with multiple concurrent medications: A comparison of methods and an application in multidrug-resistant tuberculosis

    Arman Alam Siddique, Mireille E Schnitzer, Asma Bahamyirou, Guanbo Wang, Timothy H Holtz, Giovanni B Migliori, Giovanni Sotgiu, Neel R Gandhi, Mario H Vargas, Dick Menzies, and Andrea Benedetti. Causal inference with multiple concurrent medications: A comparison of methods and...

    2019

  20. [30]

    Transcatheter versus surgical aortic-valve replacement in high-risk patients

    Craig R Smith, Martin B Leon, Michael J Mack, D Craig Miller, Jeffrey W Moses, Lars G Svensson, E Murat Tuzcu, John G Webb, Gregory P Fontana, Raj R Makkar, et al. Transcatheter versus surgical aortic-valve replacement in high-risk patients. New England Journal of Medicine, 36...

    2011

  21. [31]

    Soenksen, Yu Ma, Cynthia Zeng, Leonard Boussioux, Kimberly Villalobos Carballo, Liangyuan Na, Holly M

    Luis R. Soenksen, Yu Ma, Cynthia Zeng, Leonard Boussioux, Kimberly Villalobos Carballo, Liangyuan Na, Holly M. Wiberg, Michael L. Li, Ignacio Fuentes, and Dimitris Bertsimas. Integrated multimodal artificial intelligence framework for healthcare applications. NPJ Digital Medic...

    2022

  22. [32]

    Dropout: a simple way to prevent neural networks from overfitting

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The journal of machine learning research, 15 0 (1): 0 1929--1958, 2014

    1929

  23. [33]

    Learning prescriptive R e LU networks

    Wei Sun and Asterios Tsiourvas. Learning prescriptive R e LU networks. In Andreas Krause, Emma Brunskill, Kyunghyun Cho, Barbara Engelhardt, Sivan Sabato, and Jonathan Scarlett (eds.), Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedi...

    2023

  24. [34]

    Estimation and inference of heterogeneous treatment effects using random forests

    Stefan Wager and Susan Athey. Estimation and inference of heterogeneous treatment effects using random forests. Journal of the American Statistical Association, 113 0 (523): 0 1228--1242, 2018. doi:10.1080/01621459.2017.1319839. URL https://doi.org/10.1080/01621459.2017.1319839

    2018

  25. [35]

    Webb and David A

    John G. Webb and David A. Wood. Current status of transcatheter aortic valve replacement. Journal of the American College of Cardiology, 60 0 (6): 0 483--492, 2012. doi:10.1016/j.jacc.2012.01.071. URL https://www.jacc.org/doi/abs/10.1016/j.jacc.2012.01.071

    2012 doi

  26. [36]

    Offline multi-action policy learning: Generalization and optimization

    Zhengyuan Zhou, Susan Athey, and Stefan Wager. Offline multi-action policy learning: Generalization and optimization. Operations Research, 71 0 (1): 0 148–183, 2023. doi:10.1287/opre.2022.2271

    2023

  27. [37]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.